The Contek MC34118 voice switched speaker-phone
integrated circuit incorporates the necessary amplifiers,
attenuators, level detectors and control algorithm to form the
heart of a high quality hands-free speaker-phone system. It
includes a microphone amplifier with adjustable gain and
mute control. Transmit and Receiveattenuators which
operate in a complementary manner, level detectors at both
input and output of both attenuators, and background noise
monitors for both the transmit and receive channels. A dial
tone detector prevents the dial tone from being attenuated by
the receive background noise monitor circuit. Also two line
driver amplifiers which can be used to form a hybrid network
in conjunction with an external coupling transformer. A highpass filter can be used to filter out 60Hz noise in the receive
channel, or for other filtering functions.A chip disable pin
permits powers down the entire circuit to converse power on
long loops where loop current is at a minimum. The Contek
MC34118 may be operated from a power supply, or it can be
powered from the telephone line, requiring typically 5mA. The
Contek MC34118 can be interfaced directly to TIP and
RING( through a coupling transformer) for stand-alone
operation, or it can be used in conjunction with a handset
speech network and/or other features of a feature phone.
FEATURES
*Improved attenuator gain range: 52dB between Transmit
and Receive
*Low voltage operation for line-powered applications(3~6.5V)
*4-point signal sensing for improved sensitivity
*Background noise monitors for both Transmit and Receive
paths
SOP-28
DIP-28
*Microphone amplifier gain set by external
Resistors-Mute function included
*Chip disable for active/standby operation
*On board filter pinned-out for user defined
function
*Dial tone detector to inhibit receive idle mode
during dial tone presence
ABSOLUTE MAXIMUM RATINGS(Ta=25 C,Voltages referred to pin 22)
PARAMETERSYMBOLVALUEUNIT
Supply VoltageVcc-1~7V
Voltage at Pin 3V3-1~Vcc+1V
Voltage at Pin 12( mute)V12-1~Vcc+1V
Voltage at Pin 13( VLC)V13-1~Vcc+0.5V
Voltage at Pin 9, Pin 21, and Pin 2V9, 21, 2-0.5~Vcc+0.5V
Storage temperatureTstr-65~150C
Supply VoltageV43.5~6.5V
Voltage at Pin 3V30~VccV
Voltage at Pin 12( MUTE)V120~VccV
Voltage at Pin 13( VLC)V130.3*VB~VBV
IVB Current( Pin 15)IVB500mA
Attenuator Input Signal Voltage at Pin 9, Pin 21V9, V21350Mvrms
Load Current
@RXO, TXO(Pin8, Pin 22)
@MCO(Pin 10)
@HTO-, HTO+(Pin 6, Pin5)
Ambient Operating TemperatureTopr-20~60C
1FOFilter Output. Output impedance is less than 50 ohms.
2FIFilter Input. Input impedance is greater than 1 M ohms.
3CDChip Disable. A logic low (<0.8V) sets normal operation. A logic high (>2V) disable the IC to converse
4VccA supply voltage of +2.8V to +6.5V is required at 5mA. As Vcc falls from 3.5V to 2.8V, an AGC circuit
5HTO+ Output of the second hybrid amplifier. The gain is internally set at -1 to provide a differential output, in
6HTO- Output of the first hybrid amplifier. The gain of the amplifier is set by external resistors.
7HTIInput and summing node for the first hybrid amplifier. DC lever is about equal to VB.
8TXO Output of the transmit attenuator. DC level is about VB.
9TXIInput of the transmit attenuator. Maximum signal level is 350 mVrms. Input impedance is 10KW.
10MCO Output of the microphone amplifier. The gain of the amplifier is set by external resistors.
11MCI Input of the summing node of the microphone. DC lever is VB.
12MUT Mute input. A logic low (<0.8V) sets normal operation. A logic high (>2V) mutes the microphone
13VLC Volume control input. When VLC=VB, the receive attenuator is at maximum gain when in the receive
14CTThe RC at this pin sets the response time for the circuit to switch modes.
15VBN output voltage=Vcc/2. This voltage is a system AC ground and biases the volume control. A filter
16CPT The RC at this pin sets the time constant for the transmit background side.
17TLI2 Input to the transmit level detector on the microphone/ speaker side.
18TLO2 Output to the transmit level detector on the microphone/ speaker side, and input to the transmit
19RLO2 Output of the receive level detector on the microphone/ speaker side
20RLI2 Input to the receive level detector on the microphone/ speaker side
21RXIInput to the receive attenuator and dial tone detector. Maximum input level is 350mVrms. Input
22RXO Output of the receive attenuator. DC level is VB.
23TLI1 Input to the transmit level detector on the line side.
24TLO1 Output to the transmit level detector on the line side, and input to the transmit background monitor.
25RLO1 Output of the receive level detector on the line side.
26RLI1 Input to the receive level detector on the line side.
27CPR The RC at this pin sets the time constant for the receive background monitor.
28GND Ground pin for the entire IC.
power. Input impedance is nominally 90 K ohms.
reduces the receive attenuator gain by 25dB( when in the receive mode).
conjunction with HTO-, to the hybrid transformer.
amplifier without affecting the rest of the circuit. Input impedance is 90 K ohms.
mode. When VLC=0.3dB, the receive gain is down 35dB. Does not effect the transmit modes.
Vcc=6.5V, CD=2V
CD Input ResistanceRCDVcc=VCD=6.5V5090kW
CD Input Voltage(High)VCDH2VccV
CD Input VoltageVCDL0.8V
VB Output VoltageVBVcc=3.5V
VB Output ResistanceROVBIVB=1mA400W
VB Power Supply Rejection RatioPSRRCVB=220mF,
ATTENUATORS
Receive Attenuator Gain(f=1kHz, VLC =VB)
Rx Mode, RXI=150mVrmsGRXVcc=5V468dB
Rx Mode, RXI=150mVrmsGRXVcc=3.5V468dB
Gain ChangeDGRX1Vcc=3.5V vs.
AGC Gain ChangeDGRX2Vcc=3.5V vs.
Idle Mode, RXI=150mVrmsGRXI-22-20-17dB
Rx to Tx Mode RangeDGRX3495254dB
Volume Control RangeVCR2735dB
RXO DC VoltageVRXORx ModeVBV
DRXO DC VoltageDVRXORx to Tx Mode+-10+-150mV
RXO High VoltageVRXOHIout=-1mA,
RXO Low VoltageVRXOLIout=+1mA,
RXI Input ResistanceRRXIRXI<350mVrs71014kW
TRANSMIT ATTENUATOR GAIN(f=1kHz)
Tx Mode, TXI=150mVrmsGTX468 dB
Idle Mode, TXI=150mVrmsGTXI-22-20-17dB
Tx to Rx Mode RangeDGTXI495254dB
TXO DC VoltageVTXOTxVBVDTXO DC VoltageDVTXOTx to Rx Mode+-30+-150mV
TXO High VoltageVTXOHIout=-1mA,
HTO- to HTO+ OffsetHBVOSRf=51kW-30030mV
Open Loop GainAVOLHHTI to HTO-,
Gain BandwidthGBWH1MHz
Closed Loop GainAVCLHHTO- to HTO+-0.3500.35dB
Input Bias Current(@HTI)IBH-30nA
HTO- High VoltageVHT-HIout=-5mA3.7V
HTO- Low VoltageVHT-LIout=+5mA250mV
HTO+ High VoltageVHT+HIout=-5mA3.7V
HTO+ Low VoltageVHT+LIout=+5mA450mV
DistortionTHDM300Hz<f<10kHz0.3%
LEVEL DETECTORS and BACKGROUND NOISE MONITORS
Transmit-Receive Switching
Threshold( Ratio of Current at
RLI1+RLI2 to 20mA at TLI1+TLI2 to
Switch from Tx to Rx)
Source Current at RLO1, RLO2,
TLO1, TLO2
Sink Current at RLO1, RLO2,
TLO1, TLO2
CPR, CPT Output ResistanceRCPIout=1.5mA35W
CPR, CPT Leakage CurrentICPLK-0.2mA
Filter
Voltage Offset at FOFOVOSVFO-VB, 220kW
FO Sink CurrentIFO150260400mA
FI Bias CurrentIFI-50nA
SYSTEM Distortion(f=1kHz)
Rx ModeTHDRfrom FI to RXO,
Tx ModeTHD`from MCI to HTO-
I
TH0.811.2mA
ILSO-2mA
ILSK4mA
CONDITIONS
from VB to FI
FO connected to
RXI
/HTO+, includes
TX attenuator
MINTYPMAXUNIT
-200-900mV
0.53%
0.83%
Contek
TEMPERATURE PARAMETERS
PARAMETERTYP VALUE
Vcc Supply Current(CD=0.8V)5mA-0.3%/ C
Vcc Supply Current(CD=2V)400mA-0.4%/ C
VB Output Voltage(Vcc=5V)2.1V+0.8%/ C
Attenuator Gain(Max Gain)60.0008dB/ C
Attenuator Gain(Max Attenuation)-46dB0.004dB/ C
Attenuator Input resistance10kW0.6%/ C
Dial Tone Detector Threshold15mV20mV/ C
CT Source, Sink Current+-60mA-0.15%/ C
Microphone, Hybid Amplifier Offset0mV+-4mV/ C
Transmit-Receive Switching Threshold1+-0.02%/ C
Sink Current at RLO1, RLO2, TLO1, TLO24mA-10nA/ C
Closed Loop Gain( HTO- to HTO+)0dB0.001%/ C
(Ta=25 C
)
FUNCTIONAL DESCRIPTION
INTRODUCTION
The fundamental difference between operation of a speakerphone and a handset is that of half-duplex versus fullduplex. The handset is full duplex since conversation can occur in bothdirections(transmit and receive)
simultaneously. A speakerphone has higher gain levels in both paths, and attempting to converse full duplex results in
oscillatory problems due to the loop that exists within the system. The loop is formed by the receive and transmit
paths, the hybrid, and the acoustic coupling(speaker to microphone). The only practical and economical solution used
to date is to design the speakerphone to function in a half duplex mode- ie. only one person speaks at a time, while
the other listens. To achieve this requires a circuit which can detect who is talking, switch on the appropriate
path( transmit or receive), and switch off( attenuate) the other path. In this way, the loop gain is maintained less than
unity, when the talkers exchange function, the circuit must quickly detect this, and switch the circuit appropriately. By
providing speech level detectors, the circuit operates in a hands-free mode, eliminating the need for a push-to-talk
switch. The handset, by the way, has the same loop as the speakerphone. But since the gains are considerably lower,
and since the acoustic coupling from the ear piece to the mouthpiece is almost non-existent ( the receiver is normally
held against a person s ear), oscillations do not occur. The ContekMC34118 provides the necessary level detectors,
attenuators, and switching control for a properly operating speakerphone. The detection sensitivity and timing are
externally controllable. Additionally, the Contek MC34118 provides background noise monitors which make the circuit
insensitive to room and line noise, hybrid amplifier for interfacing to Tip and Ring, the microphone amplifier, and other
associated functions. Please refer to the Block Diagram when reading the following sections.
ATTENUATORS
The transmit and receive attenuators are complementary in function, ie. when one is at maximum gain(+6dB), the
other is at maximum attenuation(-46dB), and vice versa. They are never both fully on or both fully off. The sum of their
gains remains constant(within a nominal error band of +-0.1dB) at a typical value of -40dB( see Figure 10). Their
purpose is to control the transmit and receive paths to provide the half-duplex operation required in speakerphone.
The attenuators are non-inverting, and have a -3dB(from max gain) frequency of ~100kHz. The input impedance of
each attenuator( TXI and RXI) is nominally 10k W, and the input signal should be limited to 350 mVrms( 990mvp-p) to
prevent distortion. That maximum recommended input signal is independent of control setting. The diode clamp on
the inputs limits the input swing, and therefore the maximum negative output swing. This is the reason for VRXOL and
VTXOL specification being defined as they are in the Electrical Characteristics. The output impedance is < 10 W until
the output current limit( typically 2.5mA) is reached.
TYP CHANGEUNIT
The attenuators are controlled by the signal output of the Control Block, which is measurable at the CT pin ( pin 14).
When the CT pin is at +240mV with respect to VB, the circuit is in the receivemode(receive attenuator is at +6dB).
The circuit is in an idle mode when the CT voltage is equal to VB causing theattenuators gains to be halfway
between their fully on and fully off positions(-20 dB each). Monitoring the CT voltage(with respect to VB) is the most
direct method of monitoring the circuit s mode. The inputs to the Control Block are 7: 2 from the comparators operated
by the level detectors, 2 from the background noise monitors, the volume control, the dial-tone detector, and the AGC
circuit. These 7 inputs are described below.
LEVEL DETECTORS
There are four level detectors-two on the receive side and two on the transmit side. Refer to Figure 3-the terms in
parentheses form one system, and the other terms form the second system. Each level detector is a high amplifier
with back-to-back diodes in the feedback path, resulting in non-linear gain, which permits operation over a wide
dynamic range of speech levels. The sensitivity of each level detector is determined by the external resistor and
capacitor at each input ( TLI1, TLI2, RLI1, and RLI2). Each output charges an external capacitor through a diode and
limiting resistor, thus providing a DC representation of the input AC signal level. The outputs have a quick rise
time( determined by the capacitor and an internal 350 W resistor), and a slow decay time set by an internal current
source and the capacitor. The capacitors on the four outputs should have the same value (+-10%) to prevent timing
problems. Referring to the Block Diagram, on the receive side, one leveldetector(RLI1) is at the receive input
receiving the same signal as at Tip and Ring, and the other(RLI2) is at the output of the speaker amplifier. On the
transmit side, one level detector(TLI2) is at the microphone amplifier, while the other(TLI1) is at the hybrid output.
Outputs RLO1 and TLO1 feed a comparator, the output of which goes to theattenuator Control Block. Likewise,
outputs RLO2 feed a second comparator which also to the attenuator Control Block. The truth table for the effects of
the level detectors on the Control Block is given in the section describing the Control Block.
Contek
BACKGROUND NOISE MONITORS
The purpose of the background noise monitors is to distinguish speech ( which consists of bursts) from background
noise( a relatively constant signal level). There are two background noise monitors- one for the receive path and one
for the transmit path. Referring to Figure 3, the receive background noise monitor is operated on by the RLI1-RLO1
level detector, while the transmit background noise monitor is operated on by the TLI2-TLO2 level detector. They
monitor the background noise by storing a DC voltage representative of the respective noise levels in capacitors at
CPR and CPT. The voltage at these pins have slow rise times ( determined by the external RC), but fast decay times.
If the signal at RLI1( or TLI2) changes slowly, the voltage at CPR( or CPT) will remain more positive than the voltage
at the non-inverting input of the monitors output comparator. When speech is present, the voltage on the noninverting input of the comparator will rise quicker than the voltage at the inverting input ( due to the burst characteristic
of speech), causing its output to change. This output is sensed by the attenuator Control Block. The 36 mV offset at
the comparator s input keeps the comparator from changing state unless the speech level exceeds the background
noise by ~4dB. The time constant of the externalRC( ~4.7seconds) determines the response time to background
noise variations.
The volume control input at VLC(pin 13), is sensed as a voltage with respect to VB. The volume control affects the
attenuators only in the receive mode. It has no effect in the idle or transmit modes. When in the receive mode, the
gain of the receive attenuator will be +6dB, and the gain of the transmit attenuator will be -46dB, only when VLC is
equal to VB. As VLC is reduced below VB, the gain of the receive attenuator is reduced, and the gain of the transmit
attenuator is increased, such that their sum remains constant. Changing the voltage at VLC changes at CT ( see the
attenuator Control Block section), which in turn controls the attenuators. The volume control setting does not affect
the maximum attenuator input signal at which noticeable distortion occurs. The bias current at VLC is typically 60 nA
out of the pin, and does not vary significantly with the VLC voltage or with VCC.
DIAL TONE DETECTOR
The dial tone detector is acomparator with one side connected to the receiveinput(RXI) and the other input
connected to VB with a 15mV offset( see Figure 4). If the circuit is in the receive mode, and the incoming signal is
greater than 15Mv(10 mVrms). The comparator s output will change, disabling the receive idle mode. The receive
attenuator will then be at a setting determined solely by the volume control.
The purpose of this circuit is to prevent the dial tone ( which would be considered as continuous noise) from fading
away as the circuit would have the tendency to switch to the idle mode. By disabling the receive idle mode, the dial
tone remains at the normally expected full level.
AGC
The AGC circuit affects the circuit only in the receive mode, and only whenn the supply voltage( Vcc) is less than 3.5
volts. As Vcc falls below 3.5 volts ., the transmit path attenuation changes such that the sum of the transmit and
receive gains remains constant. The purpose of this feature is to reduce the power) and current) used by the speaker
when a line-powered speakerphone is connected to a long line, where the available power is limited. By reducing the
speaker power, the voltage sag at Vcc is controlled, preventing possible erratic operation.
ATTENUATOR CONTROL BLOCK
The Attenuator control block has the seven inputs described below:
*The output of the comparator operated by RLO2 and TLO2 ( microphone/ speaker side), designated C1.
*The output of the comparator operated by RLO1 and TLO1 ( Tip/ Ring side), designated C2.
*The output of the transmit background noise monitor, designated C3.
*The output of the receive background noise monitor, designated C4.
* The volume control
* The dial tone detector
* The AGC circuit
The single output of the control block controls the two attenuators. The effect of C1~C4 is as follows:
1. Transmit means the transmit attenuator is fully on( +6dB). And the receive attenuator is at max. attenuation
( -46dB).
2. Receive means both attenuators are controlled by the volume control. At max. volume the receive attenuator is
fully on( +6dB), and the transmit attenuator is at max.attenuation( -46dB).
3. Fast Idle means both transmit and receive speech are present in approximately equal levels. The attenuators
are quickly switched(30 ms) to idle until one speech level dominates the other.
4. Slow Idle means speech has ceased in both transmit and receive paths. The attenuators are then slowly
switched(1 second) to the idle mode.
5. Switching to the full transmit or receive from any other mode is at the fast rate(~30 ms).
A summary of the truth table is as follows:
1. The circuit will switch to transmit if: (A) both transmit level detectors sense signals relative to therespective
receive level detectors( TLI1 versus RLI1, TLI2 versus RLI2) and (B) the transmit background noise monitor indicates
the presence of speech.
2. The circuit will switch to receive if: (A) both receive level detectors sense higher signal levels relative to the
respective transmit level detectors and (B) the receive background noise monitor indicates the presence of speech.
3. The circuit will switch to the fast Idle mode if the level detectors disagree on the relative strengths of the signal
levels, and at least one of the background noise monitor indicates speech. For example, referring to the Figure 2, if
there is sufficient signal at the microphone amp.Output(TLI2) to over-ride the speaker signal(RLI2), and there is
sufficient signal at the receive input(RLI1) to over-ride the signal at the hybrid output(TLI1), and either or both
background noise monitors indicates speech, then the circuit will be in the fast idle mode. Two conditions which can
cause the fast idle mode to occur are : (A) when both talkers are attempting to gain control of the system by talking at
the same time. And (B) when one talker is in a very noisy environment, forcing the other to continually over-ride that
noise level. In general, the fast idle mode will occur infrequently.
4. The circuit will switch to the slow idle mode when: (A) both talkers are quiet ( no speech present) or (B) when one
talker s speech level is continuously over-ride by noise at the other speaker s location. The time required to switch the
circuit between transmit, receive, fast idle, and slow idle is determined in part by the components at the CT pin( pin14),
(see the section on switching timers for a more complete explanation of the switching time components A schematic
of the CT circuitry is shown in Figure 5, and operates as follows:
*RT is typically 120kW, and CT is typically 5 mF.
*To switch to the receive mode, I1 is turned on ( I2 is off), charging the external capacitor to +240mV above VB( An
internal clamp prevents further charging of the capacitor).
*To switch to the transmit mode, I2 is turned on( I1 is off), bringing down the voltage on the capacitor to -240mV with
respect VB.
*To switch to idle quickly(fast idle), the current sources are turned off, and the internal 2k W resistor is switched in,
discharging the capacitor to VB with a time constant= 2kW xCT.
*To switch to idle slowly(slow idle), the current sources are turned off, the switch at the 2k W resistor is open, and the
capacitor discharges to VB through the external resistor RT with a time constant=RT x CT.
The microphone amplifier ( Pins 10, 11) has the non-inverting input internally connected to VB while the inverting
input and the output are pinned out. Unlike most op-amps., the amplifier has an all- NPN output stage, which
maximizes phase margin and gain bandwidth. This feature ensures stability at gains less than unity, as wide range of
reactive loads. The open loop gain is typically 80Db(f<100Hz), and the gain bandwidth is typically 1MHz. The
maximum p-p output swing is typically 1 volt. less than Vcc with an output impedance of <10 W until current limiting is
reached( typically 1.5mA). Input bias current at MCI is typically 40nA out of the pin. The muting function( Pin 12),
when activated, will reduce the gain of the amplifier to ~-39dB( with RMI=5.1k W ) by shorting the range of ground and
Vcc. If the mute function is not used, the pin should be grounded.
HYBRID AMPLIFIER
The two hybrid amplifiers ( at HTO+, HTO-, and HTI), in conjunction with an external transformer, provide the twoto-four wire converter for interfacing to the telephone line. The gain of the first amplifier( HTI to HTO-) is set by
external resistors( gain = -RHF/RHI in Block Diagram), and its output drives the second amplifier, the gain of which is
internally set at -1. Unlike most op-amps ., the amplifiers have an all- NPN output stage, which maximizes phase
margin and gain-bandwidth. This feature ensures stability at gains less than unity, as with a wide range of reactive
loads. The open loop gain of the first amplifier is typically 80dB, and the gain-bandwidth of each amplifier is ~1MHz.
The maximum p-p output swing is typically 1.2 volts. less than Vcc with an output impedance of <10 W until current
limiting is reached( typically 8mA). The output current capability is guaranteed to be a minimum of 5mA. The bias
current at HTI is typically 30nA out of the pin. The connections to the coupling transformer are shown in the Block
Diagram. The block labeled ZBal is the balancing network necessary to match the line impedance.
FILTER
The operation of the filter circuit is determined by the external components. The circuit within the Contek MC34118,
from pins FI to FO is a butter with a high inputimpedance(>1MW), and a low output impedance(<50 W). The
configuration of the external components determines whether the circuit is a high-pass filter( as shown in Block
Diagram), a low-pass filter, or a band-pass filter. As a high-pass filter, with the components shown in Figure 6, the
filter will keep out 60Hz( and 120Hz) hum which can be picked up by the external telephone lines. As a low-pass filter,
with the components shown in Figure 7, it can be used to roll off the high end frequencies in the receive circuit, which
aids in protecting against acoustic feedback problems, with an appropriate choice of an input coupling capacitor to the
low-pass filter, a band pass filter is formed.
The power supply voltage at Vcc( Pin 4) is to be between 3.5 and 6.5 volts.for normal operation, with reduced
operation possible down to 2.8 volts. The output voltage at VB ( Pin 15) is ~(Vcc-0.7)/2, and provides the AC ground
for the system. The output impedance at VB is ~400 W and in conjunction with the external capacitor at VB, forms a
low pass filter for power supply rejection. The choice of capacitor is application dependent based on whether the
circuit is powered by the telephone line or power supply. Since VB biases the microphone and hybrid amplifiers, the
amount of supply rejection at their outputs is directly related to the rejection at VB, as well as their respective gains.
The Chip Disable(Pin 3) permits powering down the IC to conserve power and/or for muting purposes. With CD<0.8
volts., normal operation is in effect. With CD>2 volts. and <Vcc, the IC is powered down. In the powered down mode,
the microphone and the hybrid amplifiers are disabled, and their outputs go to a high impedance state. Additionally,
the bias is removed from the level detectors. The bias is not removed from the filter( Pin 1, 2), the attenuators( Pin 8, 9,
21, 22), or from Pins 13, 14, and 15( the attenuators are disabled, however, and will not pass a signal impedance at
CD is typically 90kW, has a threshold of ~1.5 volts., and the voltage at this pin must be kept within the range of ground
and Vcc). If CD is not used, the pin should be grounded.